Dr Agnieszka Karlińska

Agnieszka Karlińska conducts research in automatic text processing and analysis, situated at the intersection of computer science and computational linguistics, with occasional forays into digital humanities and computational social sciences. In her Ph.D. thesis, she examined gender bias in forensic psychiatric assessment. At NASK, she is involved in both developing tools for detecting harmful content, such as hate speech, and building safe and reliable large language models. Her work focuses particularly on evaluating LLMs’ tendency to generate toxic and discriminatory content and mitigating biases at various stages of model development, with a special emphasis on data selection and alignment processes.